What is Mcp Llm

mcp-llm is a command-line interface (CLI) tool designed for interacting with Anthropic’s Claude model through MCP (Model Context Protocol) servers.

Use cases

Use cases for mcp-llm include conducting interactive conversations with users, answering specific queries, and integrating with applications that require real-time language understanding.

How to use

To use mcp-llm, clone the repository, install it with ‘uv sync’, and create a JSON configuration file to define your MCP servers. You can then start interactive chats or process single queries using various command options.

Key features

Key features include the ability to connect to multiple MCP servers, interactive chat with Claude, real-time streaming of responses, support for both interactive and single-query modes, and rich terminal output with clear formatting.

Where to use

mcp-llm can be used in fields such as AI research, customer support automation, interactive applications, and any domain requiring natural language processing capabilities.

MCP Client

A CLI tool for interacting with Anthropic’s Claude model using MCP (Model Context Protocol) servers.

Features

• Connect to multiple MCP servers defined in a configuration file
• Interactive chat with Claude using available MCP tools
• Stream Claude’s responses in real-time
• Support for both interactive and single-query modes
• Rich terminal output with clear formatting

Installation

# Clone the repository
git clone https://github.com/lingster/mcp-llm.git
cd mcp-llm

# Install with uv
uv sync 

Configuration

Create a JSON configuration file that defines your MCP servers. For example:

{
  "mcpServers": {
    "brave-search": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-brave-search"
      ],
      "env": {
        "BRAVE_API_KEY": "your-api-key-here"
      }
    }
  }
}

Save this file to one of these locations:

• ~/.config/mcp-client/config.json
• ~/.mcp-client.json
• ./mcp-client.json
• ./config.json

Alternatively, you can specify a custom config file with the --config option.

Usage

Interactive Chat

# Start interactive chat with all configured servers
uv run mcpllm chat

# Connect to specific servers only
uv run mcpllm chat --server brave-search 

# Use a custom configuration file
uv run mcpllm chat --config /path/to/config.json

# Customize Claude model and parameters
uv run mcpllm chat --model claude-3-5-sonnet-20241022 --temperature 0.8 --max-tokens 8192

Single-Query Mode

# Process a single query and exit
uv run mcpllm chat --query "Is it raining in London?"

# tool use
uv run mcpllm chat --server brave-search --query "Where's wally?"

List Configured Servers

# Show all configured servers
mcpllm servers

Environment Variables

• ANTHROPIC_API_KEY: Your Anthropic API key (required)

You can also create a .env file in your current directory:

ANTHROPIC_API_KEY=your-api-key-here

License

MIT